Agricultural greenhouses, as the core component of modern facility agriculture, exhibit significant demand response potential due to their complex energy consumption characteristics. A greenhouse power demand response strategy based on crop growth safety constraints has been proposed. Firstly, the types of loads within the greenhouse are categorized, distinguishing between shiftable and interruptible loads. Then, by characterizing the nonlinear relationship between environmental parameters and electricity consumption, an electricity load model required for crop growth is established, and a greenhouse power demand response model is constructed with the objective of minimizing operating costs. To address the issue of traditional black hole algorithms easily falling into local optima, an adaptive crossover mutation mechanism and dynamic inertia weight strategy are introduced to develop an improved black hole optimization algorithm. Simulation results using a typical tomato greenhouse as a case study demonstrate that the proposed method reduces daily load and electricity purchase costs while ensuring normal crop growth. The research results provide an economical solution for demand-side response in the agrcultural greenhouse sector.